Holy ABCs! The Impact of Religion on Attitudes about Education Policies*

Holy ABCs! The Impact of Religion on Attitudes about Education Policies* Melissa Deckman, Washington College Objective. To examine the impact of religion on attitudes about three controversial education policies: creationism, school prayer, and vouchers. Methods. Using a unique, national survey of school board candidates that I undertook in 1998, I use regression analysis to examine which factors, especially religion, explain support for these three policies. Results. This study finds that conservative Christians are more likely to support such policies than mainline Protestants. Additionally, church attendance appears to have an amplifying effect on evangelicals with respect to support for these issues. By contract, those candidates with non-judeo-christian religious identifications have much lower levels of support for creationism and school prayer. However, conservative political ideology remains the strongest predictor of support for creationism, prayer in school, and vouchers. Conclusion. Religious beliefs, sometimes enhanced by church attendance, can have a powerful, direct effect on attitudes about creationism, school prayer, and vouchers among school board candidates. Demonstrating such a link becomes important when considering that school board members play a large role in shaping local education policy. The American public routinely claims that one of its top priorities is education. The public schools in this nation have no shortage of critics, chief among them Christian Right leaders, whose activism in education politics demonstrates that public education is an arena in which church-state issues are often contested. Further, many parents not just conservative Christians are calling for the government to help them pay for tuition at private, religious institutions through vouchers and tax credits. Despite the salience of education to both Christian Right activists and the public more generally, few studies have looked at the impact of religious beliefs and behavior on education attitudes. This article examines the impact of religion on three education issues that have drawn recent media attention: school prayer, creationism, and vouchers. Data for this study are drawn from a unique, national survey of school board candidates. In recent years, school *Direct all correspondence to Melissa Deckman, Washington College, Department of Political Science, 300 Washington Ave., Chestertown, MD 21620 <melissa.deckman @washcoll.edu>. All data and coding materials are available from the author for purposes of replication. The author would like to acknowledge Laura Langbein, Sue Crawford, Paul Djupe, and two anonymous reviewers for helpful advice. SOCIAL SCIENCE QUARTERLY, Volume 83, Number 2, June 2002 2002 by the Southwestern Social Science Association

Holy ABCs! 473 board elections involving conservative Christian candidates and Christian Right organizations such as the Christian Coalition have made national headlines and have heightened debate about the proper role of religion in public schools (Deckman, 1999). Analyzing the attitudes of school board candidates not only offers us a close-up examination of what potential board members think about these controversial subjects, but it also allows us to determine to what degree religion causes these opinions. Examining religion as a causal factor in attitudes about education policy is important for several reasons. Previous research establishes a strong link between religion and attitudes about other social and public policy issues. With respect to attitudes about public schools more generally, research shows that religiously conservative individuals are most likely to think that public schools are hostile to their moral and spiritual values (Sikkink, 1999). Religion s ability to influence opinions about social and political issues, combined with the salience of education in American politics today, merits an examination of how religion impacts public opinion with respect to three education policies: creationism, school prayer, and vouchers. All three issues have sparked national arguments in recent years and show how complicated, and controversial, the intersection of religion and politics can be in the United States. The debate over how students should be taught about the origins of life is not a new one in American history (Nelkin, 1982). Conservative Christian organizations argue that evolution is anathema to biblical accounts of creation and that the teaching of it can lead to dangerous societal consequences. Renewed discussions about whether students should be taught evolution, creationism, or both emerged on a national scale after a decision by the Kansas State School Board in August 1999 attempted to suppress the instruction of evolution in that state s public schools. 1 In other school districts, conservative Christian groups often fight for the inclusion of creationism in social science classes. School prayer is another perennial topic in which church and state issues collide. Conservative Christian leaders have tested the limits of the 1962 Supreme Court decision that banned statesponsored school prayer through their support of alternatives such as moments of silence. Conservative Christian leaders routinely call for a constitutional amendment that would protect religious expression such as prayer in public schools (Segers, 1997). While Christian Right organizations and other conservative religious leaders have long led fights supporting the inclusion of creationism and school prayer in public classrooms, support for vouchers among national leaders has found a broader base, particularly within the Republican party. Voucher or school choice programs involve 1 Under the leadership of its Christian Right board members, the board made a decision that limited the teaching of evolution in public schools (Benen, 1999). Less than a year later, several of the Christian board members lost their bids to be reelected to the state board, and the board later rescinded the decision.

474 Social Science Quarterly the government (local, state, or federal) giving parents direct financial aid to help defray the cost of tuition at private schools. Voucher supporters argue that such programs will push public schools to improve because public schools will be forced to compete for students with private schools. Voucher opponents look on these programs as defraying much-needed money out of the public schools and point to potential constitutional conflicts between church and state, as most private schools are religious in nature. Indeed, Catholic schools and evangelical Christian academies both stand to gain huge influxes of students and money if such programs are implemented. Conservative Christian leaders have called for greater support for creationism, prayer in school, and voucher programs. Yet it remains to be seen how the religious beliefs of school board candidates impact their attitudes on these three issues. This study uses a national survey of candidates for school board those individuals largely responsible for shaping local education policy to illustrate how religion affects opinions about creationism, school prayer, and vouchers. Methods A cross-sectional, national survey of recent school board candidates drawn from a random sample of school districts was conducted for this study in 1998. There are approximately 15,000 school boards in the United States, 98 percent of which are elected boards. The U.S. Department of Education maintains a list of each school district in the nation. This list, with a few modifications, was the sampling frame from which 300 school districts were randomly drawn as the primary sampling unit. 2 By using the entire list of school districts as its initial sampling frame, the study can speak quite confidently about the generalizability of the survey results. Because the majority of school districts had relatively few numbers of students compared with very large districts, the sample was divided into three groups based on size to account for this disparity: small school districts (less than 2,000 students, N = 6,627), medium school districts (2,000 9,999, N = 6,275), and large school districts (10,000 or more students, N = 690). The 300 districts were then drawn randomly in a proportionate manner, with a total of 147 small school districts, 137 medium school districts, and 16 large school districts selected. 3 The districts were contacted and asked to provide lists of all can- 2 The Directory of Public and Elementary and Secondary Education Agencies (1993 1994) provided the list of all 14,881 U.S. schools districts. Appointed school boards, school boards that did not report student size, special school districts (such as schools for the blind), and school districts with less than 25 students were excluded from analysis. This brought the final primary sampling frame to 13,582. 3 The school districts represent a diverse cross-section of the nation, coming from 39 different states. Regionally, the random selection of districts is as follows: midwest (41 percent), west (27 percent), northeast (19.5 percent), and south (12.5 percent). Finally, 65 percent of

Holy ABCs! 475 TABLE 1 Identification with, Church Attendance by Religious Tradition School board candidates identifying with religious tradition Attends church weekly Attends church monthly Attends church several times per year Attends church rarely, if ever Conservative Traditions Fundamentalist Christian 16% 63% 12% 17% 8% Pentecostal/charismatic 3 61 17 17 6 Evangelical Christian 15 82 14 1 3 Traditional Catholic 17 66 18 9 7 Liberal Traditions Mainline Christian 24 50 22 14 14 Liberal Christian 11 28 14 25 34 Conservative/reform Jew 3 12 24 47 18 Progressive Catholic 13 58 18 11 13 Other Traditions Ethical or secular humanist 2 8 15 23 54 Agnostic 4 8 4 88 Religious non-christian 2 22 33 44 NOTE: N = 578. didates who ran in the most recent school board election held in their particular districts. More than 91 percent of the districts responded, yielding a final list of 1,220 former school board candidates, who represent the secondary sampling unit. After two waves of mailings, I received 671 usable responses from former candidates (response rate 55 percent). 4 Measures I use regression analysis (ordinary least squares) to determine the impact of religion on school board candidates attitudes about creationism, school prayer, and vouchers while controlling for other factors. The dependent variables are derived from Likert scales. The survey asks former school board candidates whether they strongly support, support, oppose, strongly oppose, or are neutral/undecided about the three topics under study: voucher sys- the school districts selected come from rural areas, 27 percent from suburban areas, and 7 percent from urban areas. 4 Among respondents, 62 percent won their elections. The generalizability of the survey is supported by the fact that almost an equal percentage of nonrespondents 59 percent also won their races.

476 Social Science Quarterly tems (including tuition tax credits) to pay for tuition at private schools; opening class with a prayer; and creationism being taught as an alternative theory to evolution in science courses (1 = strongly oppose, 5 = strongly support). Religious Variables Religious tradition refers to a group of religious communities that share a set of beliefs that generates a distinctive worldview (Kellstedt et al., 1996:176). Survey respondents are asked which of any religious terms reflected their religious beliefs, including fundamentalist Christian, Pentecostal or charismatic Christian, evangelical Christian, mainline Christian, liberal Christian, conservative/reform Jew, 5 traditional Catholic, progressive Catholic, ethical or secular humanist, agnostic, and religious non-christian. These categories, which are not mutually exclusive, are included in the analysis. Table 1 contains a breakdown of religious tradition as selfidentified by the candidates in the first column. Not included in the table, or in subsequent analyses, are 93 individuals who declined to answer the question or who checked that none of the above terms reflected their religious beliefs (reducing the sample size to 578). Conservative religious categories include fundamentalist Christian, Pentecostal/charismatic Christian, and evangelical Christian. All three traditions share in common several core evangelical religious beliefs: (1) a belief in the divinity of Jesus Christ, established through a salvation/conversion experience (being born again ), (2) an acceptance of the authority of the Bible, and (3) a commitment to bearing public witness about one s faith (Shibley, 1998). Fundamentalists are unique from other evangelicals in that they believe that the Bible is not just authoritative, but inerrant. Historically, fundamentalists have also typically rejected modernism in theology and culture, a posture that has meant a more separatist identity from other evangelicals (Smith, 1998). Politically, it has meant that while both fundamentalists and nonfundamentalist evangelicals tend to embrace personally conservative positions on moral and social issues, fundamentalists are sometimes more likely to translate personal values into demands for legal action, although the two groups tend to be similar in terms of partisanship and voting behavior (Jelen, 1987). Though related to evangelicals, Pentecostals and charismatics are unique in that they emphasize the power one receives through the baptism of the Holy Spirit (Smidt, 1989:54), which they believe empowers them to engage in supernatural acts such as speaking in tongues and 5 Only one respondent identified himself/herself as an Orthodox Jew. As this individual also identified himself/herself as a reform/conservative Jew, he/she was grouped in the latter category.

Holy ABCs! 477 faith healing. 6 Like evangelicals and fundamentalists, Pentecostals and charismatics have more conservative views on political and social issues than the electorate at large, although they are less inclined to be mobilized politically than are other evangelicals (Smidt, 1989). The liberal religious categories under consideration include mainline Protestants, liberal Protestants, and conservative/reform Jews. Jews form their own distinctive religious communities set apart from the other Christian traditions. However, Jews share with mainline and liberal Protestants more moderate positions on homosexuality, abortion, and the role of women in society and the church. Historically, mainline and liberal Protestants have been more willing than evangelicals (and particularly fundamentalists) to accommodate cultural changes and modernism, such as an acceptance of evolution theory as being compatible with their religious faiths (McKinney, 1998; Smith, 1998). This acceptance of science, for example, stems from mainliners belief that the Bible is more open to interpretation, with an emphasis on symbolism. Mainliners also accept different sources of religious truth aside from the Bible, and rather than emphasize a need for evangelism, emphasize a Social Gospel of helping those individuals in need, following Christ s example in the Bible (Warner, 1988). In many ways, the Catholic Church incorporates elements from and shares similar social beliefs as both evangelical and mainline Protestants. For instance, although the Catholic Church tends to be conservative on abortion and homosexuality, it typically takes more liberal stands on social justice issues such as the death penalty or poverty. At the risk of oversimplification, this split personality manifests itself in a split among practicing Catholics into those who are more traditional and those who are more progressive. I consider in this analysis two separate Catholic categories. As the labels imply, traditional Catholics such as Pat Buchanan are generally more conservative in their politics, while progressive Catholics Senator Ted Kennedy is an example are more liberal. Finally, respondents were also given the opportunity to check ethical or secular humanist, agnostic, or religious non-christian as alternatives to the Judeo-Christian groups. Secular humanism is rooted in the idea that values and virtues are human-made rather than ordained by a higher being. Secular humanism, then, is distinguished by placing moral action in the welfare of humanity rather than in fulfilling the will of God (Honderich, 1995). To be an agnostic, however, is not to reject a higher power per se, but rather to admit an uncertainty about the existence of God (Honderich, 1995). The religious non-christian category is meant to capture those individuals who have a religion set apart from Christianity. 6 Pentecostal Christians and charismatic Christians generally share the same beliefs about the baptism of the Holy Spirit. Unlike pentecostals, however, charismatic Christians are located within already established religious traditions such as Lutherans, Methodists, and even Catholics (see Smidt, 1988).

478 Social Science Quarterly Although each of the religious categories has its own distinct religious beliefs, survey respondents could check all the terms that they believed applied to them. The multiple-identification option recognizes that several of the religious identities are compatible and may better capture an individual s religious beliefs (Wilcox, Jelen, and Leege, 1993). However, the majority of respondents 86 percent (N = 493) selected only one category. Moreover, those who checked more than one category did so in ways that are, for the most part, theologically consistent. For example, 23 individuals checked that they were both fundamentalist and evangelical. 7 The implications of such multiple identifications will be discussed shortly. In addition to religious beliefs, I also examine the impact of church attendance on attitudes about creationism, school prayer, and vouchers. Past studies have shown that regular church attendance is positively associated with political behavior such as voting and has been linked to some policy opinions, particularly conservative social and political attitudes (Wald, Kellstedt, and Leege, 1993). Table 1 also breaks down church attendance by candidate s religious self-identification. How church attendance will impact school board candidates attitudes about these three issues is unclear. On the one hand, increased church attendance could be positively linked to support for creationism, school prayer, and vouchers. One reason is that frequent church attendance might make candidates more comfortable with prayer and concepts such as creationism, which will make them more likely to support these policies in the schools. Catholics, for example, might be more exposed to pleas by church leaders to support voucher programs, which could translate into support for such programs. However, frequent church attendance might reinforce negative attitudes about these issues if the church or synagogue in question opposes these issues. Hypotheses With respect to school prayer, I should find that an evangelical, fundamentalist, pentecostal/charismatic, or traditional Catholic orientation among school board candidates would lead them to support the inclusion of prayer in the classroom. Among candidates from other, more liberal religious traditions, and those from non-christian or humanist traditions, I should find less support for school prayer. I should also find that conservative evangelicalism particularly fundamentalism is significantly related to 7 Two individual respondents, however, checked that they were both fundamentalist and liberal Christians, which seems theologically counter-intuitive. But the term liberal might be contaminated with political meaning, so these individuals remain in the study (see Wilcox, Jelen, and Leege, 1993). There are also individuals who classify themselves as fundamentalists and traditional Catholics (3) and progressive Catholics (2), respectively. Again, they might be using the term fundamentalist as a doctrinal or congregational measure rather than a theological one. Given the possible ambiguous interpretation of that term, I keep them in the data set.

Holy ABCs! 479 support for the inclusion of creationism in public schools given the authority that these traditions place on the Bible. Given that more liberal religious traditions are willing to accept evolution theory as not inconsistent with biblical beliefs, school board candidates from these traditions should not be likely to support the teaching of creationism in public schools, nor will humanists, agnostics, or religious non-christians. With voucher programs, school board candidates from conservative religious traditions should again show support. However, progressive Catholics as well as traditional Catholics might also support voucher programs given that these programs will directly benefit Catholic schools. Among the remaining liberal religious groups, Jews in particular should be opposed to vouchers, given that Jewish groups are at the forefront of supporting a separation of church and state. Given that vouchers stand to benefit Christian schools, we should find that humanists, agnostics, and religious non-christians will not support them. The relationship between church attendance behavior and attitudes about these three education issues is not as theoretically compelling as the relationship between these issues and religious beliefs. We cannot necessarily assume that high levels of church involvement automatically translate into increased support for the three education policies under study. Given past research on church involvement and its link to conservative political attitudes (Wald, Kellstedt, and Leege, 1993), however, I predict that increased church attendance will be positively related to support for these issues. Other Control Variables In addition to these religious variables, I also use standard control variables such as political ideology (1 = very liberal to 7 = very conservative), education (1 = high school education or less, 4 = graduate degree), income (1 = less than $25,000, 5 = more than $100,000), gender (1 = female, 0 = male), and race (1 = white, 0 = nonwhite). To test for geographical and community differences, I also control for region (three dummies for living in the south, midwest, and west, with the northeast being the reference variable) and urbanicity (two dummies for whether the respondent resides in a rural or suburban community; urban community is the reference category). Findings Table 2 lists a breakdown of support for creationism, school prayer, and vouchers among school board candidates by religious tradition. The differences in attitudes between the more conservative traditions and the more liberal and non-christian traditions are pronounced and consistent through all three categories as hypothesized, and many differences are significant using the chi-square statistic. Support for the teaching of creationism, for example, is strongly rooted among school board candidates who are funda-

480 Social Science Quarterly TABLE 2 Religious Tradition by Support for Education Issues Creationism being taught as an alternative theory to evolution in science courses Opening class with a prayer Voucher systems to pay for tuition at private schools Conservative Traditions Fundamentalist Christian 74%*** 69%*** 43%*** Pentecostal/charismatic a 65*** 56 33 Evangelical Christian 73*** 55* 41* Traditional Catholic 44*** 52** 34* Liberal Traditions Mainline Christian 36 32 20 Liberal Christian 22** 23* 9** Conservative/reform 13 13*** 13* Jew a Progressive Catholic 25** 30 25 Other Traditions Ethical/secular humanist a 8*** 8*** 15 Agnostic 8*** 11*** 26 Religious non-christian a 0 0* 0 NOTE: Percentages show combined totals of those candidates who support and strongly support the issues. Chi-square tests were run to test significant differences between individuals from each religious tradition versus individuals who are not part of that tradition. N = 578. *p < 0.05; **p < 0.01; ***p <0.001. a The chi-square statistics for Jews, religious non-christians, and ethical humanists, however, should be interpreted with caution. All three groups have a small sample size (N < 20), which leads to some cells having fewer than five cases. mentalist, pentecostal, and evangelical Christians. While the majority of neither group of Catholic candidates supports teaching creationism, traditional Catholics are almost twice as likely to support it than are progressive Catholics. The split between progressive and traditional Catholics is even more apparent with respect to support for prayer in school: 52 percent of traditional Catholic school board candidates support school prayer compared with 30 percent of progressive Catholics. Support for school prayer is highest among the conservative groups, particularly fundamentalists, although support for prayer is not as high as support for creationism among these same groups. Candidates from more liberal religious traditions and other traditions are hesitant to support either creationism or school prayer. Strong support for vouchers does not exist among candidates, regardless of religious tradition. Not a single majority of school board candidates from the different religious traditions indicates a willingness to support vouchers. However, the same sort of differences that existed between candidates from more conservative and liberal religious traditions appear again. The highest percentage of support comes from fundamentalist school board candidates,

Holy ABCs! 481 followed by evangelical school board candidates. Support for vouchers among Jews, liberal Christians, and other traditions is particularly low. The lack of support among Catholic school board candidates is surprising, given Catholic leaders support for vouchers. However, support for voucher programs might be limited among school board candidates more generally, given that their goal is to help govern public not private schools. Not shown in Table 2 are the isolated effects among the multiple religious identifiers those candidates who checked more than one religious category to describe their religious beliefs. As most of the multiple identifiers are consistent with their religious groupings (i.e., they tend to check multiple categories within the conservative or liberal/other religious traditions), I examined the level of support among individuals who identified with more than one conservative tradition (N = 32) and those who identified with more than one liberal or other tradition (N = 16). Results for both groups of individuals show levels of support in the expected direction, although the levels of support among multiple conservative identifiers 90 percent support for creationism, 75 percent support for school prayer, and 56 percent support for vouchers are noticeably higher than for individuals who identify with just one of the four conservative religious categories. Among multiple liberal-religious/other-religious identifiers, 13 percent support creationism, 19 percent support prayer, and 13 percent support vouchers. The interaction effect for those individuals who identify with more than one conservative religious tradition suggests some support for Wilcox, Jelen, and Leege s (1993) assertion that multiple identification with conservative religious traditions might be a reflection of high levels of religiosity. For these school board candidates, the saliency of their religious beliefs appears to lead to exceptionally high levels of support for the policy positions espoused by their conservative religious traditions. Examining the raw percentages of support given to these three issues reveals some not-so-surprising trends with respect to the relationship between religious tradition and support for creationism, school prayer, and vouchers. The next step is to use regression analysis to see how these relationships hold once controls, including church attendance, are introduced. I created three initial regression models with support for creationism, school prayer, and vouchers acting as dependent variables. For the major religious independent variables, I have opted to use dummy variables since the vast majority of respondents opted to select just one category to describe their religious beliefs. The omitted category is mainline Protestant. 8 There have been a few modifications to some religious variables used in the model. First, given that 8 Mainline Protestants are the omitted category, in part, because their mean level support scores for creationism and prayer in school are closest to the mean level of support among all respondents. In the case of vouchers, their level of support is the second closest (behind progressive Catholics) to the overall mean level. However, given that Catholics traditional or progressive may have the most theoretically interesting findings with respect to support for vouchers, I opted to keep mainliners as the omitted category.

482 Social Science Quarterly most of those survey respondents who identified themselves as pentecostal or charismatic Christians also identify themselves with another religious tradition (12 out of 18), I have dropped this group from the regression. Second, I have combined those religious categories that are not Judeo- Christian ethical humanists, agnostics, and religious non-christians into one category, in part because Table 2 indicates that they tend to share the same beliefs on all three of the issues under study. Finally, I also introduce interaction terms between each of the religious self-identifications and church attendance to account for the fact that religious traditions often place different emphasis on the value of church attendance (see Wald and Smidt, 1993). I show the three initial models with the different dependent variables in Table 3, as well as the same models after including the interaction for church attendance and religious tradition. Turning to the first regression model support for creationism among school board candidates I find, as expected, that being a fundamentalist or evangelical Christian makes a school board candidate significantly more likely than a mainline Protestant candidate to support creationism. The same cannot be said, however, for those candidates who are traditional Catholics. I find that while the sign for liberal Christians, Jews, progressive Catholics, and individuals from other religious traditions indicates that these groups are less likely than mainliners to support creationism, the relationship is significant only for the other category. Church attendance, meanwhile, is positively related to support for creationism. Among control variables, ideology and education are significant. Ideology, in particular, shines as a strong predicator of support, given that it has the largest beta coefficient in the model. The more politically conservative a school board candidate, the more likely he or she is to support creationism. Alternatively, the more education a candidate has, the less likely he or she is to support creationism. The addition of interaction terms to the model in the next column which allows me to test whether church attendance has an especially powerful effect on individuals from particular religious traditions reveals that church attendance does condition the impact of religious tradition for evangelical Christians and liberal Christians. The more frequently evangelicals attend church, the more likely they are to support the instruction of creationism in public schools. Increased church attendance among liberal Christians, however, results in less support for creationism instruction.

Holy ABCs! 483 TABLE 3 OLS Regression: Impact of Religion on Education Attitudes Regression Results Standardized (Beta) Coefficients Reported Support for Creationism Creationism Support for Prayer Prayer Support for Vouchers Vouchers Religious Variables Fundamentalist 0.181*** 0.019 0.167*** 0.146 0.102* 0.059 Evangelical 0.150*** 0.076 0.048 0.007 0.096* 0.060 Trad. Catholic 0.013 0.164 0.149*** 0.190 0.147*** 0.160 Liberal Christian 0.002 0.165 0.041 0.100 0.036 0.167 Jewish 0.051 0.086 0.093* 0.205* 0.052 0.023 Prog. Catholic 0.038 0.001 0.020 0.065 0.090* 0.230 Other 0.117** 0.183* 0.181*** 0.272*** 0.096* 0.034 Church Attendance 0.108** 0.070 0.056 0.035 0.043 0.117 Personal Factors Ideology 0.255*** 0.232*** 0.220*** 0.205*** 0.316*** 0.312*** Education 0.148*** 0.123*** 0.036 0.018 0.030 0.019 Gender 0.029 0.040 0.089* 0.098*** 0.048 0.060 Race 0.061 0.035 0.123*** 0.104** 0.097** 0.086* Regional, Community Differences Midwest 0.031 0.035 0.072 0.072 0.032 0.032 West 0.025 0.022 0.105* 0.110* 0.532 0.029 South 0.018 0.012 0.132** 0.128** 0.038 0.042 Rural 0.013 0.014 0.086 0.085 0.121 0.137 Suburban 0.019 0.020 0.088 0.091 0.068 0.076 Interaction, Religion by Church Attendance Fun*Attend 0.184 0.313*** 0.022 Evan*Attend 0.228*** 0.141*** 0.144** TCath*Attend 0.102 0.006 0.027 LibChr*Attend 0.200** 0.090 0.126 Jew*Attend 0.052 0.119 0.028 PCath*Attend 0.004 0.112 0.123 Other*Attend 0.093 0.089 0.138 R 2 0.337 0.384 0.360 0.386 0.181 0.205 NOTE: *p < 0.05; **p < 0.01; ***p < 0.001. The next column in Table 3 looks at support for school prayer among school board candidates, which differs from support for creationism in several respects. First, while fundamentalist school board candidates are still significantly more likely than mainliners to support prayer in school, evan-

484 Social Science Quarterly gelical candidates are not. More interestingly, however, is that candidates who identify themselves as traditional Catholics are significantly more likely than mainliners to support school prayer. The relationship between support for school prayer and Jewish and other religious candidates works in the opposite direction: candidates from both groups are significantly more likely to oppose school prayer than mainliners. Conservative political ideology is positively associated with support for prayer among school board candidates, and again has the largest beta weight in the model. For the first time, race becomes a significant factor as well: nonwhite school board candidates are more likely to support school prayer than white candidates. Gender emerges in this model as significant, with men more likely than women to support prayer in school, while education is unrelated to support for school prayer. Finally, significant regional differences emerge among the school board candidates, with candidates from the south being more likely, and western candidates less likely, to support school prayer than candidates from the northeast. Unlike the creationism model, church attendance is not significantly related to support for school prayer (although the direction of the beta coefficient is positive). However, the interaction terms reveal a strong, positive relationship among fundamentalist and evangelical Christians and church attendance, which suggests that church attendance has an amplifying effect on the socialization of candidates from these two traditions that is not present among candidates from other faiths. Finally, the last model examines support for vouchers among school board candidates. As predicted, those candidates who identify themselves as traditional Catholics are significantly more likely than mainliners to support voucher plans. The model also finds that progressive Catholics are significantly more likely to support vouchers than are mainline Protestant candidates. The same is also true among fundamentalist and evangelical Christians. What is surprising, however, is that candidates from other religious traditions are significantly more likely than mainline Protestants to support voucher programs. This finding might reflect the fact that vouchers, although they stand to directly benefit religious schools, are less religious in nature than policies that promote creationism or prayer. The only two groups that appear to have less support for vouchers than mainline Protestants are liberal Christians and Jews, although the relationship is not statistically significant. Holding conservative political views once again appears as a highly significant factor when examining attitudes about vouchers. Similar to the school prayer model is the finding that nonwhite candidates are more likely than whites to support vouchers. This finding is not surprising considering that African-American church leaders often support vouchers as an experimental approach to assist inner-city parents in getting their children out of sub-par public schools. Support for vouchers does not seem to be significantly related to education, gender, region, or community type. Neither does church attendance, although the negative direction of the coeffi-

Holy ABCs! 485 cient might reflect a tendency among some churches to preach that less government interference in church-run institutions, such as schools, might be better. Clearly, however, this tendency does not apply to the churches attended by evangelical Christian school board candidates, as the introduction of interaction terms reveals that increased church attendance among evangelicals leads to more support for vouchers. Discussion This study finds strong, independent relationships between religion and attitudes regarding support for creationism, school prayer, and vouchers, even while controlling for political views, race, gender, and regional and community differences. Not surprisingly, differences fell largely between the conservative and liberal religious camps. A fundamentalist belief among school board candidates is significantly related to support for all three controversial policies in the initial models. Evangelicals are more likely than mainliners to support creationism and vouchers, while traditional Catholics are more likely to support prayer in school and vouchers. Church attendance is positively related to support for creationism, although it appears to amplify support among evangelical school board candidates for all three policies (and for fundamentalists when it comes to school prayer). With one exception, church attendance tends to be significant only when it interacts with the more conservative religious traditions, which is not surprising considering that evangelical traditions place more emphasis on and allow for more opportunity for group worship (Wald, Kellstedt, and Leege, 1993). Support for vouchers among evangelical Christian and Roman Catholic leaders might help to explain why fundamentalist, evangelical, traditional and progressive Catholic identifiers are significantly more likely than mainliners to support vouchers. Harder to explain is why those candidates who identify with other, non-christian traditions are also more likely to support vouchers than mainliners in the model. Support for vouchers may be driven less by religious tradition and more by politics. The major factor related to support for all three issues is political ideology the more conservative the candidate, the more likely he or she is to support creationism, school prayer, and vouchers. 9 Race also emerges as a factor for school prayer and support for vouchers. Nonetheless, findings from this analysis demonstrate that religious beliefs and practices can have a powerful direct effect on attitudes about various education policies. Demonstrating such a link between religion and education attitudes among school 9 Moreover, ideology might hide some indirect effects of religion. Although space limits a more thorough examination of this point, path analyses show that many of the religious traditions have a substantial impact when regressed upon ideology as a dependent variable. In fact, all religious traditions have a significant relationship when regressed on ideology in the expected directions, with the exception of Jews and mainline Protestants.

486 Social Science Quarterly board candidates becomes especially important when one considers that at least some of these candidates will have a hand in shaping education policies on local school boards. Future research might concentrate on exploring similar links between religion and the attitudes of state-level school board members and officials involved with establishing education policies. REFERENCES Benen, Steve. 1999. Evolution Evasion. Church & State. October:4 8. Deckman, Melissa M. 1999. Christian Soldiers on Local Battlefields: Campaigning for Control of America s School Boards. Ph.D. dissertation. American University. Honderich, Ted, ed. 1995. Oxford Companion to Philosophy. New York: Oxford University Press. Jelen, Ted G. 1987. The Effects of Religious Separatism on White Protestants in the 1984 Election. Sociological Analysis 48(1):30 45. Kellstedt, Lyman A., et al. 1996. Grasping the Essentials: The Social Embodiment of Religion and Political Behavior. Pp. 174 92 in John C. Green, James L. Guth, Corwin E. Smidt, and Lyman A. Kellstedt, eds., Religion and the Culture Wars: Dispatches from the Front. Lanham, Md.: Rowman & Littlefield. Leege, David C. 1993. Religion and Politics in Theoretical Perspective. Pp. 3 25 in David C. Leege and Lyman A. Kellstedt, eds., Rediscovering the Religious Factor in American Politics. Armonk, N.Y.: M.E. Sharpe. McKinney, William. 1998. Mainline Protestantism 2000. Annals of the American Academy of Political and Social Sciences 558(July):57 66. Nelkin, Dorothy. 1982. The Creation Controversy: Science or Scripture in the Schools. New York: Norton. Segers, Mary C. 1997. The Religious Equality Amendment and Voluntary School Prayer. Pp. 191 218 in Jo Renee Formicola and Hubert Morken, eds., Everson Revisited: Religion, Education, and Law at the Crossroads. Lanham, Md.: Rowman & Littlefield. Shibley, Mark A. 1998. Contemporary Evangelicals: Born-Again and World Affirming. Annals of the American Academy of Political and Social Sciences 558(July):67 87. Sikkink, David. 1999. The Social Sources of Alienation from Public Schools. Social Forces 78(1):51 86. Smidt, Corwin. 1989. Praise the Lord Politics Politics: A Comparative Analysis of the Social Characteristics and Political Views of Evangelical and Charismatic Christians. Sociological Analysis 30:53 72. Smith, Christian. 1998. American Evangelicalism: Embattled and Thriving. Chicago, Ill.: University of Chicago Press. Wald, Kenneth D., Lyman A. Kellstedt, and David C. Leege. 1993. Church Involvement and Political Behavior. Pp. 121 38 in David C. Leege and Lyman A. Kellstedt, eds., Rediscovering the Religious Factor in American Politics. Armonk, N.Y.: M.E. Sharpe. Wald, Kenneth D., and Corwin E. Smidt. 1993. Measurement Strategies in the Study of Religion and Politics. Pp. 26 52 in David C. Leege and Lyman A. Kellstedt, eds., Rediscovering the Religious Factor in American Politics. Armonk, N.Y.: M.E. Sharpe.

Holy ABCs! 487 Warner, R. Stephen. 1988. New Wine in Old Wineskins: Evangelicals and Liberals in a Small- Town Church. Berkeley, Cal.: University of California Press. Wilcox, Clyde, Ted Jelen, and David C. Leege. 1993. Religious Group Identification: Toward a Cognitive Theory of Religious Mobilization. Pp. 72 99 in David C. Leege and Lyman A. Kellstedt, eds., Rediscovering the Religious Factor in American Politics. Armonk, N.Y.: M.E. Sharpe.